The obesity epidemic is a continuing and growing major global multi-scale problem. Designing appropriate policies and interventions has been challenging since obesity is a complex problem, crossing the following six scales: genetic, physiological, individual, group/social network, physical (built) environment, and societal. The overall goal of this proposed project is to develop the Virtual Populations for Obesity Prevention (VPOP), a software platform that can generate an agent-based model encompassing the six obesity- relevant scales of any metropolitan area that can help decision makers to design, evaluate, and test proposed (or existing) obesity interventions and policies. VPOP will bring multiple innovations by (1) being the first model to bring together and integrate the six different scales that affect obesity; (2) including novel representations of numerous pathways and relationships; (3) leading to new insights and targets for obesity-control policies and interventions; (4) being grounded in an unprecedented breadth and depth of real- world multi-scale obesity-related data; (5) heavily involving decision makers in multi-scale model development to maximize policy-relevancy and translation of VPOP results into useful action; (6) developing new ways of representing and visualizing multi-scale results; and (7) transforming obesity-related data collection and decision making. Our multi-disciplinary team is led by Global Obesity Prevention Center (GOPC), which focuses on developing and implementing multi-scale systems science approaches, methods, and tools to address obesity, and brings together experts from the Pittsburgh Supercomputing Center (PSC)/ Carnegie Mellon University (CMU), Cornell, and the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). The proposed VPOP and our participation in Interagency Modeling and Analysis Group (IMAG), and Multi-scale Modeling Consortium (MSM) activities would substantially leverage our existing GOPC and PSC/CMU resources and efforts. This includes extensive field studies to provide a large and broad data set to help populate, calibrate, and validate the VPOP and forming a Stakeholder Working Group to guide VPOP development, testing, and implementation. Specific Aim 1 will develop VPOP, a platform that can generate a geospatially explicit computational model representing the six obesity-relevant scales for any metropolitan area. Specific Aim 2 will entail utilizing VPOP to generate multi-scale simulation models of two sample metropolitan areas (the Baltimore Metropolitan Statistical Area and New York City) to use to identify the key drivers of obesity in children and adults across the six different scales and determine which factors may be maximally sensitive to specific programs and policies. For Specific Aim 3, we will translate the VPOP- generated models to decision-making by working with key stakeholders, such as city policy makers, health and planning department leadership, healthcare administrators, clinicians, and third-party payers to test and optimize a specific set of obesity-control policies and interventions.